On 21 Jan 1999, Jeroen Pluimers writes:
> What is the best way to install ssh and get a secure connection from
> a (don't shiver!) Windows 95/NT machine to log on to a linux
> machine?
(1) Install SSH on Linux from the RPMs on ftp.replay.com.
(2) Pick up the free (as in beer, not as in speech) SSH32 client for
Win32 from http://www.doc.ic.ac.uk/~ci2/ssh and also the associated
crypto library version 1.10. Install those on the Win32 box. Put the
crypl110.dll file in the same place you put ssh32.exe. It then pretty
much just works.
(3) If you need good terminal emulation under Win32 (say to run Emacs
well), get TeraTerm and set up a tunnel in SSH32 so when you telnet to
localhost (the Win32 box) you get an encrypted conection to port 23
(telnet) on the Linux host.
You can also spend money on easier to install Win32 SSH clients such
as the one from F-Secure or the one from Van Dyke Technologies. The
Van Dyke one has much better terminal emulation than the freeware
SSH32 one, but... well, I'm cheap, so I use SSH32 and then use
TeraTerm (open source terminal emulator for Win32) over an SSH tunnel
to get decent terminal emulation.
I tend to carry a diskette with SSH32 on it around so I can get at
some of the systems here when at other people's homes from their Win9x
boxes :-)
> The error-message appeared on the console while RPM was updating
> linuxconf (without anyone being logged in on it - strange!).
Not necessarily strange at all. Check /etc/syslog.conf for an entry
containing /dev/console. Probably syslog was told to send certain
kinds of error messages to the console, and it did. That's normal
behaviour. You can edit your /etc/syslog.conf to ensure that all
error messages you want to see remotely are logged to files in
/var/log, or even sent to the logged in user root whereever s/he may
be, if you prefer that.
> Because of this, the console kept some linuxconf processes running.
This I don't understand. The console is just a device. It has no
controlling influence over other processes.
Wait a second... *Linuxconf* processes? Why were these running at
all? Did you run Linuxconf and then (while it was running) try to
upgrade it while it was running?? I've certainly not tried that... it
would be more conventional, and probably safer (!) to avoid running
linuxconf at the exact same time as you upgrade it <grin>.
> RPM refused to continue and got so confused that it actually had
> installed TWO versions of linuxconf at one time!
OK... I can't think how, unless you specify a non-default database
location to RPM. That sounds like an RPM bug to me. Again it would
be really valuable to duplicate this rather just ignore it.
> The only way to get it right, was running RPM at the console and
> manually press OK on these error messages. Only then RPM would
> continue installing 1.13r12 and removing 1.12r7. After that,
> upgrading to the right managerpm solved the other problems.
Well, I'm glad you solved it for yourself, but it would be much better
if we could duplicate this and so solve it for everyone. Error
messages on the console (from rpm? from linuxconf?) that you have to
press OK to get to disappear sounds unusual... not syslogd in that
case... maybe RPM itself? I just got a copy of Maximum RPM but
haven't read it yet. Maybe I should!
Any RPM experts out there care to comment on this, and on how to
prevent RPM (or was it Linuxconf... I'm confused now!) from sticking
error messages and OK prompts on the console (text mode console I
presume) when noone is logged in there and rpm is run from some other
pty? It does sound like, um, less-than-optimal behaviour :-)
Jonathan
--
Jonathan Marsden | Internet: [EMAIL PROTECTED] | Making electronic
1849 N. Wabash Ave.| Phone: +1 (909) 794 1151 | communications work
Redlands, CA 92374 | FAX: +1 (909) 794 3016 | reliably for Christian
USA | http://www.xc.org/jonathan | missions worldwide
---
You are currently subscribed to linuxconf as: [[email protected]]
To unsubscribe, forward this message to [EMAIL PROTECTED]
